Liver fibrosis is the principal feature of the injury caused by chronic liver disease and determines the major clinical events that lead to liver-related deaths. For this reason alone, an accurate assessment of fibrosis is vital to the management of patients with liver disease.

Measuring the extent of liver disease is also a significant factor when considering whether to use treatment, to assess the response to therapy and to make other important decisions related to progression of fibrosis, such as screening for hepatocellular carcinoma and varices.

For 60 years, liver biopsy has been regarded as the gold standard diagnostic for assessing the progression of fibrosis in chronic hepatitis C patients. However, despite its longstanding utility, liver biopsy has some significantly negative features. First, patients often resist undergoing liver biopsy due to the discomfort resulting from its invasiveness. There is also some risk to the patient of experiencing an adverse event from liver biopsy. In addition to these negative features, there is a sampling error of at least 24% due to inadequate liver specimen length or fragmentation. Finally, there are inconsistencies in the interpretation of liver biopsy results because of errors on the part of one or more observers of the specimens.

As a result of these drawbacks to liver biopsy, interest continues to grow in new, non-invasive methods of assessing fibrosis, including biochemical markers, biomarkers, and new imaging techniques.

FibroScan and FibroTest

One such advance in the field is FibroScan, a type of ultrasound machine that uses transient elastography to measure liver stiffness. The device reports a value that is measured in kilopascals (kPa). This value can be extrapolated to a fibrosis score.

FibroTest (aka FibroTest-ActiTest) is another non-invasive diagnostic for assessing fibrosis. Available through BioPredictive (www.biopredictive.com), FibroTest uses an algorithm to combine the results of serum tests of beta 2-macroglobulin, haptoglobulin, apolipoprotien A1, total bilirubin, gamma glutamyltranspeptidase (GGT), and alanine aminotransferase (ALT) to assess the level of fibrosis and necroinflammatory activity.

Comparison of FibroScan and FibroTest to Detect Fibrosis Progression among HCV Carriers with Normal Aminotransferases

An article published in the October 2005 issue of Hepatology by Colletta et al compared the value of FibroScan and FibroTest in HCV carriers with normal ALT levels [1]. The study evaluated 40 untreated HCV RNA positive subjects who had two liver biopsies, with a median interval of 78.5 months, during which ALT levels never exceeded 1.2 times the upper normal limit.

The study authors concluded that FibroScan yielded results that showed perfect agreement between FibroScan and liver biopsy [emphasis added—Ed].  In addition, the study concludes that the diagnostic accuracy of FibroScan was 100%. Further, the authors write, “FibroScan is superior to the FibroTest in the noninvasive identification of fibrosis, for which excess alcohol consumption in the past and high viral load represent risk factors” [2].

The glowing review by Colletta et al of the diagnostic superiority of FibroScan compared to FibroTest has now been strongly challenged by other experts in the field, specifically L Castera et al and T Poynard et al. Their contrasting opinions appear in the “Correspondence” section of the February 2006 issue of Hepatology [3,4]. A reply by Colletta et al to Castera et al and to Poynard et al also appears in the February 2006 issue of Hepatology. The major arguments of each of the three teams of experts concerning the original study by Colletta et al are summarized here.

FibroScan and FibroTest to Assess Liver Fibrosis in HCV with Normal Aminotransferases

(L Castera and others)

Although Castera et al agree with the opinion of Colletta et al that these non invasive diagnostics could “someday become an alternative” to liver biopsy in patients with persistently normal ALT levels (PNAL), the authors state they feel compelled to voice several methodological concerns about the study by Colletta et al.

Castera et al raise three major issues about the study:

1.       The stated capability of FibroScan to identify the entire spectrum of fibrosis stage in individuals with PNAL contracts sharply with the results noted in all other published studies. Due in part to the small number of patients in their study, Colletta et al need to interpret their data more cautiously  concerning the “perfect agreement” between FibroScan and liver biopsy, write Castera et al.

            The performance of FibroScan in this study shows a surprising diagnostic accuracy of 100%. A       critical issue in assessing accuracy is the cutoff value for identifying patients with significant          fibrosis. Colletta et al give no justification for the choice of their cutoff point of O.31. Using the cutoff proposed by Castera et al, the results by Colletta et al most likely would be quite             different than what they report, with lower diagnostic accuracy: sensitivity 100%, specificity             46%, diagnostic accuracy 46%, positive predictive value 50%, negative predictive value    100%).

2.       Second, Colletta et al offer no information on the proportion of patients in whom liver elasticity measurements could not be obtained.

3.       Third, Colletta et al should be cautious in asserting that FibroScan has much better correlation with liver biopsy than FibroTest. In their study, FibroTest yields a surprisingly high false positive rate, which contrasts with the results of prior published studies that show high specificity for FibroTest.

Castera et al propose avoiding the limitations of both FibroScan and FibroTest by employing an algorithm that combines both tests. When doing so, the authors found that in 97 of 100 consecutive HCV patients with normal ALT levels, concordance between FibroScan and FibroTest for significant fibrosis was 67%.

In conclusion, Castera et al write, “We believe combining FibroScan and FibroTest as a first-line noninvasive assessment of liver fibrosis might prove particularly useful in the setting of HCV carriers with normal ALT levels and should be further evaluated.”

Service d'Hépato-Gastroenterologie, Hôpital Haut Léveque, C.H.U.Bordeaux, Pessac, France

Diagnostic Value of FibroTest with Normal Serum Aminotransferases (T Poynard et al)

Colletta et al have concluded that FibroScan is superior to FibroTest.  Their study yielded “perfect” diagnostic measurements for FibroScan, but for FibroTest (at the 0.31 cutoff) only 64% sensitivity and 31% specificity for the diagnosis of advanced fibrosis.

In the observations of Poynard et al, such “poor” results have not been found for FibroTest, nor are they reflected in the results of other published studies using the services of BioPredictive, “the sole company allowed to market FibroTest.”

In a FibroTest analysis of 537 patients with chronic hepatitis C (129 with normal ALT and 408 with elevated ALT), there were no differences in area under the ROC curves between patients with normal or elevated ALT. Furthermore, the AUROC was 0.76, higher than the 43% accuracy observed by Colletta et al.

The Poynard group also analyzed prospectively the specificity of FibroTest in 954 blood donors without liver biopsy, 917 with normal ALT. Excluding the 25 patients with high risk of false positive/negative, 877 of the remaining 892 patients (98.2%) have normal FibroTest and 15 (1.8%) had FibroTest between 0.31 and 0.48 and none above. “These figures are very different than the 69% false positive of Colletta et al,” according to Poynard at al.

Poynard et al propose four possible explanations for these discrepancies:

1.       The small number of patients in the study by Colletta et al.

2.       An error in the calculation of FibroTest if the professional website (www.BioPredictive.com) was not used.

3.       The non exclusion of patients with high risk profile of false positive/negative. (These patients would have been identified by the security algorithms on the BioPredictive website).

4.       Use of the FibroTest threshold at 0.31 for F2F3F4 detection instead of 0.48 which is the recommended threshold.

Finally, in evaluating the concordance between the two tests in 70 consecutive subjects with baseline normal ALT: 60 patients (44 HCV, 5 hepatitis B, 11 others), and 10 apparently healthy volunteers, FibroScan was not applicable in 3 subjects (abdominal fat) and FibroTest in 5 (high risk profile). In the remaining 62 subjects the concordance was fair with 78% (48/62) of concordance for stage F2F3F4 when using the 0.31 threshold, and even better (82% (51/62) at the recommended threshold (0.48). These concordance rates were similar to the 77% observed by Castera et al.

In conclusion, Poynard et al write, “We think the comparisons between noninvasive markers should be performed according to professional recommendations, respecting applicability reports to exclude high risk of false negative or false positive results and in populations with sufficient sample size.”

Service dHepato-Gastroenterologie, GH Pitie-Salpetriere, Pans, France; BioPredictive, Paris, France

Colletta et al Reply to Poynard et al and Castera at al

Colletta et al dismiss two of the explanations for the discrepancies between their study and the study by Poynard et al by stating that all FibroTest values were calculated using the BioPredictive website. Further, the five serum markers used in this fibrosis index were measured following the strict requirements specified by BioPredictive, say Colletta et al.

Concerning the use of the 0.31 cutoff, Colletta et al maintain that this is the most sensitive cutoff to exclude significant fibrosis. Further, if the data are re-analyzed using the 0.48 cutoff, the diagnostic value of FibroTest in the studied patients would be: sensitivity, 21%, specificity, 81%, positive predictive value, 38%, negative predictive value, 66%.

While acknowledging that future, larger studies might reach different conclusions than theirs, Colletta et al emphasize that the selection of patients for these future studies will require the use of very strict criteria, such as those used by them (no value >1.2 times the upper normal limit in a minimum of 17 determinations, along more than 5 years), and which Poynard et al did not follow. This included the using patients with PNAL who have ALT within the normal limits on the day a liver biopsy was performed.

Regarding the objections of Castera et al related to the cutoff chosen for liver elasticity (8.7%kPa), using a cutoff that reduces the overall value of a test is not justifiable, say Colletta et al.

With regard to the second issue raised by Castera et al., the proportion of patients in whom liver elasticity measurements could not be obtained, Colletta et al reply, “Unfortunately, Castera et al neither specify why their patients could not have liver elasticity measured, nor define the HCV carriers with normal ALT they studied in terms of length of follow-up and number of ALT determinations.”

Finally, the authors question the appropriateness of Castera et al to equate the specificity of FibroTest in blood donors without hepatitis C (and no liver biopsy) to that in HCV carriers with normal ALT, “since these two populations should not be considered equivalent.”

It would appear that FibroScan has certain advantages over other diagnostic indices or predictive models based on laboratory tests in that it is completely noninvasive, provides a more direct measure of fibrosis, should not be affected by other disease states, and should theoretically be applicable to all chronic liver diseases, explain Ghany and Doo in an editorial that accompanies the study by Colletta et al [7].

Conclusion

What will be the future of the liver biopsy and the newer, non-invasive diagnostics such as FibroScan and FibroTest?

If as expected, new antivirals become available for the treatment of chronic hepatitis C, there will be an increasing need to assess fibrosis as a method of monitoring the effect of these treatments. Liver biopsy no doubt will continue to be employed in the diagnosis, grading and assessment of chronic liver disease. Yet despite a continuing role for liver biopsy, non invasive methods likely will soon become the diagnostic of choice for assessing liver fibrosis.

02/07/06

Sources

C Colletta C and others. Value of two noninvasive methods to detect progression of fibrosis among HCV carriers with normal aminotransferases. Hepatology 42(4): 838-845. October 2005.

L Castera, J Foucher, J Bertet, P Couzigou, and V de Ledinghen. FibroScan and FibroTest to assess liver fibrosis in HCV with normal aminotransferases. Hepatology 43(2): 373-374. February 2006.

T Poynard, M Munteanu, Y Ngo, M Torres, Y Benhamou, D Thabut, and V Ratziu. Diagnostic value of FibroTest with normal serum aminotransferases. Hepatology 43(2): 374-375. February 2006.

References

1.       C Colletta C and others. Value of two noninvasive methods to detect progression of fibrosis among HCV carriers with normal aminotransferases. Hepatology 42(4): 838-845. October 2005.

2.       Ibid.

3.       L Castera, J Foucher, J Bertet, P Couzigou, and V de Ledinghen. FibroScan and FibroTest to assess liver fibrosis in HCV with normal aminotransferases. Hepatology 43(2): 373-374. February 2006.

4.       T Poynard, M Munteanu, Y Ngo, M Torres, Y Benhamou, D Thabut, and V Ratziu. Diagnostic

      value of FibroTest with normal serum aminotransferases. Hepatology 43(2): 374-375.

            February 2006.

5.       C Colletta and others. Reply to L Castera et al and T Poynard et al. Hepatology 43(2): 375-376. February 2006.

6.       M G Ghany and E Doo. Assessment of liver fibrosis: Palpate, poke or pulse? (editorial). Hepatology 42(4): 759-761. February 2006.